When a School Adds a New Responder to Your Child’s Life

When your child talks with a teacher, a coach, a grandparent, a neighbor, or a friend’s parent, the adult in that conversation carries a responsibility your child does not. The adult decides when to answer, when to redirect, when to stop, and when to call you. That is not something parents have to arrange in advance; it is simply part of how adult responsibility works around children.

Generative AI changes the conversation

The AI systems your child may now use at school are not search engines. A standard search engine provides links to relevant websites after the search, and may now even summarize an answer, and stops. A generative AI system can hold a full conversation. Your child types something, the system replies, your child replies to that, and the exchange continues. The system may remember what came before, build on it, suggest where to go next, and shape each answer around the conversation already underway.

That means the concern is often not one bad answer, but the whole exchange. A conversation can travel a long way from where it began, one small step at a time.

Think about what a teacher does when a child asks an unusual question. The question may be off-topic. It may be too personal. It may ask for too much help. It may be slightly alarming. The teacher does something ordinary but important: the teacher interprets. The teacher reads the question against the child they know and decides whether it is curiosity to encourage, a boundary to redirect, or a signal that a parent, counselor, or principal needs to know.

No policy can fully do that in advance. A policy is general. Judgment makes it specific and actionable.

A good teacher also knows how much help is too much. The teacher knows when to give a small nudge to a student who is stuck, and when to hold back because the struggle is part of the learning. A teacher who simply did the work for a child would not be helping; (s)he would be failing the role. And if that happened, the failure would have an address: a person to ask, a principal above her, and a school responsible for what occurred.

For certain serious disclosures, a teacher’s response is not merely a matter of preference. Teachers operate within legal and institutional duties, including mandated reporting obligations in defined situations. They may still get things wrong, but they are part of an accountable system.

Now consider your child’s conversation with a generative AI system on a school-issued device.

The system answers. It advises. It responds individually, in real time, over an ongoing exchange. It may encourage, redirect, explain, suggest, or complete the child’s work. But unless the school has built and verified a governance structure around it, the system does not occupy the same accountable position as the adults in your child’s life. It receives the practical authority to guide the conversation, without carrying the obligations that normally come with that authority. It does not know your child and does not answer to you. It is not a mandated reporter. It is not bound by the school’s curriculum judgment in the same way a teacher is. It has no child-facing duty unless some accountable institution has made that duty real and enforceable.

The question is simple: who is responsible for how that conversation goes?‍

Filters are not the same as judgment

Your school may say the system has filters or guardrails. Filters may catch some obvious things, and that has value. But filters do not replace judgment. A filter usually reads messages against rules or categories. It may not understand the full trajectory of a conversation, especially when each individual message looks harmless on its own.

A child may reach a troubling place through ordinary words: “yes,” “what next,” “show me,” “I don’t know,” “keep going.” Those words may mean very little as isolated messages. They may mean much more as the tenth or fiftieth step in a conversation that has been drifting. A teacher standing inside that conversation could notice the drift and step in. A filter may see nothing wrong.

That is why logs matter. If the school cannot see the full sequence of the interaction — not just isolated messages, but the depth and direction of the exchange — it may not be able to reconstruct what happened. And if it cannot reconstruct what happened, it cannot answer the basic questions parents are entitled to ask: What happened? Why did it happen? Who was responsible? And what will prevent it from happening again?

The same problem appears with schoolwork. Some AI systems are designed to nudge rather than solve. Others will simply produce the answer, write the paragraph, or complete the assignment. Both may be reachable through the same device, account, browser, or platform environment. A parent who would object if a teacher did the child’s homework has the same reason to object if a school gives the child a tool that can do it, while no adult is able to see that it happened.

And that is the point: the only witness may be the child.

Parents may not be able to see what happened

Parents parent on information. You know where your child spends time. You know which adults are around them. When something new enters your child’s life, you usually assess it yourself. You get to know families, neighbors, teachers, coaches. If something feels off to you, it’s meaningful. Your experience is reasonably predictable.

With generative AI, that powerful, everyday parental method may fail. Your own experience of the system may tell you little to nothing about your child’s experience. The same tool that gives you a recipe, a packing list, or a harmless explanation may be having a very different conversation with your child. Your “it seemed fine” may be based on a conversation your child is not having.

This is one of the few things generative AI has in common with social media. The technologies are fundamentally different, but they share two parent-facing problems: your experience as the parent is not predictive of your child's experience, and you may have no meaningful visibility into what your child is actually experiencing. And unless the school technically disables access outside of direct teacher supervision, no two children are having the same experience with it. On a given evening, one child may ask a single question their teacher suggested. Another child may be having an ongoing conversation that ranges wherever it goes, where the system keeps suggesting what to ask next. In that second conversation, the judgment about what is within school policy and what is not falls to the child — a role that, in a classroom, would never be handed to a student. That is the teacher's job. And once access is given, the child may also be the only witness to what happened, left to self-report it. That is not supervision. It is a fallback when supervision has already failed.

What Has Your School Asked Your Child to Do?

So the real question is not about the technology or if GenAI belongs in school or not. It is about what the school has placed on your child. Imagine the school handed your child a book and said: you may only read chapters 2 and 4, and only at these times. Then it relied on your child to comply. That would not be meaningful supervision. Meaningful supervision would mean handing over the right chapters, or a different book entirely. That is what responsible access looks like: the adult controls what reaches the child, not the child.

Now ask whether the school has done the same with generative AI. In many schools, the device or account is mandatory for participation, including homework, assignments, teacher communication, or classroom platforms. If generative AI is accessible through that same device or account, parents should ask whether the school is controlling access technically or expecting the child to self-police it. In other words, is it the school's job to block what your child should not reach — or your child's job to simply not go there? Is your child expected to translate school policy into safe questions, in the moment, turn after turn? To judge whether each thing the system suggests is within bounds? To know if a simple response like “tell me more” or “yes” breaks policy? To be held responsible not for the question asked, but for the answer the system gave back? In a classroom, none of that would ever be handed to a student. It is the teacher's job. The question a parent can ask is simple: which of these has my child been made responsible for, and did anyone tell me?

‍ ‍

The Formal Questions to Ask Your School

Before You Ask - An explanation in “plain English”

When a school makes generative AI available to students on school-issued devices or accounts, parents do not need to become technology experts. They need clear answers to basic governance questions. If a school has added a new responder to your child’s life, parents are entitled to know who is responsible for that responder, what it can do, what it is telling your child, and what happens when the conversation goes somewhere it should not.

These questions are designed to find out:

• Which generative AI systems and features students can actually reach on school-issued devices and accounts.

• Whether the school can see and reconstruct student interactions when something goes wrong.

• Who is responsible for the parts of the conversation that filters miss, teachers cannot see, parents cannot supervise, and children cannot reasonably be expected to govern alone.

Some parents may want to send only a few questions first and wait for a response. Others may want to create a more complete written record. Both approaches are reasonable. Start with the shortest version unless you have a reason to send the full list.

When possible, ask for written answers by email and request that the questions be directed to the school or district’s legal, risk management, technology, and student services teams, not only to a classroom teacher or principal.

Start With Four Questions

This is the best starting point for most parents.

Suggested Email

Subject: Questions Regarding Student Access to Generative AI

Dear [Principal/Superintendent/Technology Office],

I am writing as a parent to better understand whether and how students can access generative AI systems on school-issued devices, school accounts, or school-approved platforms. By generative AI, I mean tools such as ChatGPT, Google Gemini, Microsoft Copilot, Khanmigo, and AI features built into platforms such as Google Docs or similar programs.

Please direct these questions to the appropriate legal, risk management, technology, and student services personnel, and please provide written responses by email.

1. Access
   What generative AI systems or features can students reasonably reach on school-issued devices or accounts, including systems that are formally approved, embedded inside approved platforms, or otherwise technically accessible?

2. Reconstruction
   If a generative AI system or feature allows free-form inputs and multi-turn interactions, can the school reconstruct the full interaction if something goes wrong, including student prompts, system outputs, system suggestions, auto-completions, and the full sequence of the exchange? If yes, please identify which systems this reconstruction is available for.

3. Accountability for AI Judgment
   A generative AI system does not exercise human judgment. But during a student interaction, it may perform functions that normally require adult judgment: answering, redirecting, refusing, suggesting, encouraging, evaluating, or completing work. Who has accepted responsibility for those system response functions, including when a harm, concern, or disputed interaction arises? 

4. Student Distress Detection and Escalation

   I have one more question, and I want to be clear up front that I'm asking about how things work in general. This is not about a specific concern with my child.

   If my child typed something into an AI tool on a school device or school account saying they were thinking about hurting themselves — the kind of statement that, said out loud to a teacher, would require the school to act — would any adult at the school ever find out it happened? If yes, who is notified, and how quickly?

Thank you,
[Parent Name]

If the school doesn’t answer the questions directly, but gives you policies or assurances, here is a guide on how to respond back.

You may also forward school responses to info@digitalchildhoodcouncil.org